参考电网奔德斯分解与协调的优化方法

发布时间：2018-05-31 03:58

本文选题：电力系统 + 参考电网　；参考：《山东大学》2015年硕士论文

【摘要】：电网规划是保证电力系统可靠运行的前提。随着可再生能源并网发电的快速发展,以及大规模电网互联的建设和市场机制的推进,电网规划面临更严峻的挑战。参考电网综合考虑了系统运行与年输电成本,且保持电力系统可靠性的一致性水平,将输电容量作为待求量,因此能够决策电网的最优输电容量。参考电网能够提供一套对现有电网进行客观评估的标准,它能对投资成本和阻塞成本进行最优量化；通过对比实际线路建设容量与参考电网相应线路容量,可以发现线路的容量裕度是否合理。考虑未来电网的发展,结合负荷预测和电源规划,以参考电网为依托,可以实现电网渐进规划策略。该策略针对参考电网对某一时间段实际电网网架结构的评估研究所发现的问题,对考虑扩建、再建的若干候选方案,研究建立动态的、考虑过程化的,在满足电网一定安全、可靠水平下的最优输电能力的优化决策模型,给出不同情景下输电供给和需求达到平衡时的值,使电网逐步趋于满足长期边际成本,即社会效益最大化下的输电能力。对于大规模电网来说,参考电网是一个复杂的优化问题。由于需要得出每条线路的最优建设容量,并且需要考虑多种事故状态的影响,优化问题的待求变量、约束条件数量庞大,求解代价大。因此,本文提出将奔德斯分解的方法应用到参考电网模型的求解中。为了符合奔德斯分解对模型的要求,本文在原始参考电网模型的目标函数中添加了考虑断线事故引起的电网操作费用,然后分别构建奔德斯主问题与子问题：将正常状态下求解最优发电调度与最优线路建设容量的优化问题作为主问题,将事故状态下操作费用最小化的优化问题作为子问题,两者通过奔德斯割关联在一起,实现原本规模庞大问题分解与协调的优化计算。根据模型的实际情况,考虑到子问题数量众多且约束存在大量冗余的特点,采取一系列缩减计算规模的策略以提高算法求解效率：利用功率分配因子可以对事故状态的潮流进行预估,在此基础上可以不经过潮流计算,实现对线路潮流与线路容量的比较分析,从而实现对线路或者约束条件关键集的筛选；提出了主导事件筛选策略,该策略显著缩减了有效预想事件的数量,从而减少了需要求解的子问题的数量,极大提高了算法整体的运行效率；提出了子问题支路潮流约束的缩减策略,在进行子问题求解时考虑只添加越限线路的潮流约束,从而减少了子问题约束条件的数量,提高了子问题求解的速度。经算例验证这些策略能显著提高运算效率。
[Abstract]:Power network planning is the premise to ensure the reliable operation of power system. With the rapid development of renewable energy grid power generation, the construction of large-scale power grid interconnection and the promotion of market mechanism, power grid planning is facing more serious challenges. Considering the system operation and the annual transmission cost and keeping the consistency level of the power system reliability, the transmission capacity is regarded as the required quantity, so the optimal transmission capacity of the power network can be determined. The reference grid can provide a set of objective evaluation criteria for the existing grid, it can optimize the investment cost and congestion cost, by comparing the actual line construction capacity with the corresponding line capacity of the reference grid, It can be found that the line capacity margin is reasonable. Considering the future development of power grid, combined with load forecasting and power supply planning, and relying on reference grid, the progressive planning strategy of power grid can be realized. This strategy aims at the problems found by the research on the evaluation of the actual grid structure in a certain period of time. For some candidate schemes considering the expansion and reconstruction, the dynamic and procedural methods are studied to meet the security of the power network. The optimal decision-making model of the optimal transmission capacity at reliable level is given, and the value of transmission supply and demand equilibrium is given under different scenarios, which makes the power network gradually meet the long-term marginal cost, that is, the transmission capacity under the maximization of social benefits. For large scale power grid, reference grid is a complex optimization problem. Due to the need to obtain the optimal construction capacity of each line and the need to consider the influence of various accident states, the number of constraints and the cost of solving the optimization problem are large. Therefore, the method of Bandes decomposition is applied to the solution of reference grid model. In order to meet the requirements of Bundes decomposition model, this paper adds the operation cost of power network to the objective function of the original reference power network model, which takes into account the power network operation cost caused by the breakage accident. Then the main problem and sub-problem are constructed respectively: the optimization problem of optimal generation scheduling and optimal line construction capacity under normal condition is taken as the main problem, and the optimization problem of minimizing the operating cost in the state of accident is taken as the sub-problem. The two methods are related to each other by Bandes' cutting to realize the optimization calculation of the decomposition and coordination of the original large scale problems. According to the actual situation of the model, considering the large number of subproblems and the large number of constraints, A series of strategies are adopted to reduce the calculation scale to improve the efficiency of the algorithm. The power allocation factor can be used to estimate the power flow of the accident state, and on this basis, the power flow can be calculated without power flow. Through the comparison and analysis of line power flow and line capacity, the selection of key sets of line or constraint conditions is realized, and a dominant event selection strategy is proposed, which significantly reduces the number of effective expected events. Thus, the number of sub-problems needed to be solved is reduced, and the overall running efficiency of the algorithm is greatly improved. Thus, the number of constraint conditions is reduced and the speed of solving subproblems is improved. Examples show that these strategies can significantly improve the computational efficiency.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TM715

【共引文献】